Abstract

BASTOS GUTIERREZ, Sara; PANIAGUA JIMENEZ, Diana  and  VARGAS MARTINEZ, Alejandro. Google Earth Engine: Evaluation of the scopes and limitations for the resolve of agricultural problems in Santa Lucía, Barva, Heredia. Rev. Geog. Amer. Central [online]. 2026, n.76, pp.313-334.  Epub Dec 12, 2025. ISSN 2215-2563.  http://dx.doi.org/10.15359/rgac.76-1.11.

Google Earth Engine (GEE) is presented as an innovative tool for agricultural management through the geospatial analysis of satellite imagery. In this study, GEE was used to classify soils at the Santa Lucía Experimental Farm (FESL), employing Sentinel-2 imagery from the Copernicus program with a resolution of 20 × 20 m throughout the year 2022. Four training classes were considered (pastures, forests, coffee, and infrastructure) using the Random Forest classifier. Additionally, a pixel-by-pixel confusion matrix was generated for both the training and validation processes. The results show an overall accuracy of 96% for the training set and 61% for the validation set, highlighting the model’s efficiency in class distinction, although with potential for improvement in distinguishing between coffee and forest classes.

Keywords : Agriculture; Google Earth Engine; land use; satellite images; Sentinel 2.

        · abstract in Spanish | Portuguese     · text in Spanish     · Spanish ( pdf )